Dedicated Servers
  • Instant
  • Custom
  • Single CPU servers
  • Dual CPU servers
  • Servers with 4th Gen CPUs
  • Servers with AMD Ryzen and Intel Core i9
  • Storage Servers
  • Servers with 10Gbps ports
  • Hosting virtualization nodes
  • GPU
  • Sale
  • VPS
  • Dedicated GPU server
  • VM with GPU
  • Tesla A100 80GB & H100 Servers
  • Sale
  • VMware and RedHat's oVirt Сlusters
  • Proxmox VE
  • Colocation
  • Colocation in the Netherlands
  • Remote smart hands
  • Services
  • Intelligent DDoS protection
  • Network equipment
  • IPv4 and IPv6 address
  • Managed servers
  • SLA packages for technical support
  • Monitoring
  • Software
  • VLAN
  • Announcing your IP or AS (BYOIP)
  • USB flash/key/flash drive
  • Traffic
  • Hardware delivery for EU data centers
  • About
  • Careers at HOSTKEY
  • Server Control Panel & API
  • Data Centers
  • Network
  • Speed test
  • Hot deals
  • Sales contact
  • Reseller program
  • Affiliate Program
  • Grants for winners
  • Grants for scientific projects and startups
  • News
  • Our blog
  • Payment terms and methods
  • Legal
  • Abuse
  • Looking Glass
  • The KYC Verification
  • Hot Deals


    Fabula - a neural network that will draw your unique avatar from your photo

    server one
    Rent GPU servers with instant deployment or a server with a custom configuration with professional-grade NVIDIA Tesla H100 / H100 80Gb or RTX A5000 / A4000 cards. GPU servers with game RTX4090 cards are also available.

    Generative neural networks for creating maps based on user requests are one of the main trends of the last year in the field of AI. Not only IT giants but also small startups are hard at work developing them. The latter have to compete with corporations with unlimited resources and robust infrastructure. How do they manage to do this? Let's take the example of Fabula, a client of the international hosting provider HOSTKEY.

    Fabula is a neural network with a focus on personalization

    Fabula was created by a team of programmers from Kazakhstan led by Ali Ozdiev. Among the many services that use generative neural networks to create pictures on demand, Fabula stands out for its focus on personalization.

    It generates customized avatars based on user-uploaded photos and text descriptions. This sets it apart from other popular neural networks that work solely on text.

    For example, upload your photo and complete it with a comment such as "Avatar in the style of Salvador Dali's paintings" or "wearing a wizard's cloak in front of a castle." The avatar can then be animated, making your cloak flutter and so on. Among the latest innovations on offer are choices of artistic styles, creating gifs and videos, as well as the ability to generate an avatar for your favorite cat or any other pet.

    How it works: you upload 5-10 photos to a telegram bot or a mobile app for iOS/Android or the project website, select a style, add comments on the details to be included in the picture, and you get a ready-made avatar that is sure to garner much attention in messengers and generate a lot of likes on social media.

    The basic functionality of Fabula is free for users, and the service is monetized through premium options. To get a free avatar created by the neural network based on user submitted photos, the users usually wait in line for 6-7 days. If you don't want to wait, you can opt for paid options and get your images within a few minutes.

    Examples of how it works

    If you upload multiple portrait photos to the service, the neural network will generate a number of differently styled avatars.

    George Clooney

    George Clooney
    George Clooney
    George Clooney
    George Clooney
    George Clooney
    George Clooney
    George Clooney
    George Clooney
    George Clooney
    George Clooney
    1 /9

    You can also add a text description, and then the neural network will give you the pictures according to your wishes.

    Portrait of a woman, unity 8k wallpaper, ultra detailed, beautiful and aesthetic, beautiful, masterpiece, top quality, ecstasy flowers, dynamic angle, the most beautiful form of chaos, elegant, vivid colors, romantic, atmospheric

    Emilia Clarke

    Emilia Clarke
    Emilia Clarke
    Emilia Clarke
    Emilia Clarke
    Emilia Clarke
    Emilia Clarke
    Emilia Clarke
    1 /7

    Launching and scaling Fabula with HOSTKEY support

    For the user of the service everything is simple and clear, but its creators had to solve a whole host of minor technical problems related to processing and storing uploaded photos, training the neural network on user data, improving algorithms for generating pictures and the data exchange speed.

    An important criterion for the success of a project like this is not only the quality of generated pictures and the speed of the service, but also the optimization of costs for its support and development. The creators of Fabula thought that it was more cost-effective to outsource the construction of the server infrastructure and turned to HOSTKEY.

    - “We do not specialize in the installation, maintenance and uninterrupted operation of equipment. At some point, we had to choose between assembling a separate team for this, or turning to a professional player who can do it better, faster and cheaper," says Ali Ozdiev. – “Renting capacity from cloud providers is very expensive, and we would not have been able to provide our users with premium service at a competitive price while working at an unsustainable loss. That's why we chose HOSTKEY, a reliable partner who provided inexpensive capacity and fast scaling when we needed it.”

    HOSTKEY has been providing server equipment rental and hosting services in European data centers for many years. One of the company's flagship products is high-performance dedicated and virtual servers, including those based on the latest generations of NVIDIA graphics cards. HOSTKEY always has servers on standby, ready to go. And the provider's own teams of engineers work in the data centers to ensure rapid deployment and hardware support.

    What the HOSTKEY team did to launch and scale Fabula:

    • Purchased the required number of NVIDIA RTX A5000 and RTX A6000 professional GPU cards.
    • Deployed the equipment in a data center in Finland, where electricity is cheaper than in other European countries, thus reducing the cost of computing power.
    • Developed and implemented a plan for commissioning new equipment based on the project development plans.

    Fabula today

    Fabula's neural network processes about 15 thousand requests daily, and their number is constantly growing. The main audience comes to the service via a bot on the Telegram messenger app. Requests to create avatars come from all over the world ranging from the US and Canada to as far as China and other Asian markets. Some of these countries have requirements to host the data in local jurisdictions. HOSTKEY, which has space in data centers all around the world, can also help in meeting these requirements.

    Ali Ozdiev:
    “The first thing that comes to mind when planning projects with unpredictable workloads is a cloud solution where you can quickly and easily manage the amount of resources you employ. But, solutions from large providers are very expensive. We opted for a cheaper and ultimately more flexible infrastructure for our avatar generation, which is built on dedicated servers. HOSTKEY's competitive pricing policy fits very well with our P&L, allowing us to keep the cost of premium services low for users, processing speed high and to maintain acceptable queue lengths for the free services. Based on our positive experience, we are now thinking about moving our storage servers to HOSTKEY sites as well."

    Fabula tomorrow

    The functionality of the service is constantly evolving. For example, the service has recently introduced a choice of avatar styles, keyword generation, and other features.

    The creators of Fabula have plans to place Android apps in third-party stores, including App Gallery.

    All of this ensures rapid audience growth in different countries. Fabula's founders expect the number of users of their services to reach 10 million by 2025.

    The HOSTKEY team is already setting up one more site in another European data center to meet the needs of this fast-growing service, further reducing the cost of the project and expanding the capacity of their neural networks. The Fabula Team is coordinating with HOSTKEY engineers on the further development of their IT infrastructure.

    The case study of the Fabula team demonstrates how you can develop your project in the field of generative neural networks without owning a huge fleet of equipment and without investing millions of dollars in server infrastructure. By choosing HOSTKEY as a partner, Fabula's creators were able to focus on developing the service's core functionality, marketing and promotion. And at the same time, they can rest assured that their services will work reliably even in times of peak load and are able to quickly scale along with their business.

    We wish the Fabula team every success and are happy to be part of our partner's success story.

    Rent GPU servers with instant deployment or a server with a custom configuration with professional-grade NVIDIA RTX 5500 / 5000 / A4000 cards. VPS with dedicated GPU cards are also available . The GPU card is dedicated to the VM and cannot be used by other clients. GPU performance in virtual machines matches GPU performance in dedicated servers.

    Other articles


    How to choose the right server with suitable CPU/GPU for your AI?

    Let's talk about the most important components that influence the choice of server for artificial intelligence.


    VPS, Website Hosting or Website Builder? Where to host a website for business?

    We have compared website hosting options, including VPS, shared hosting, and website builders.


    How AI is fighting the monopoly in sports advertising with GPUs and servers

    AI and AR technologies allow sports advertising to be customized to different audiences in real time using cloud-based GPU solutions.


    From xWiki to static-HTML. How we “transferred” documentation

    Choosing a platform for creating a portal with internal and external documentation. Migration of documents from cWiki to Material for MkDocs


    Test Build: Supermicro X13SAE-F Intel Core i9-14900KF 6.0 GHz

    Test results of a computer assembly based on the Supermicro X13SAE-F motherboard and the new Intel Core i9-14900KF processor overclockable up to 6.0 GHz.

    HOSTKEY Dedicated servers and cloud solutions Pre-configured and custom dedicated servers. AMD, Intel, GPU cards, Free DDoS protection amd 1Gbps unmetered port 30
    4.3 67 67